Make sure _NEW_POLYGONSTIPPLE state is set. A simple call case like the
following:
glEnable(GL_POLYGON_STIPPLE);
glPolygonStipple(pattern);
glClear(GL_COLOR_BUFFER_BIT); /* <---!!---> */
glRectf(...);
glFlush();
would simply make the polygon stipple fail. The glClear would fire a
draw call, it then would reset those state. So, when it about to hangle
the glRectf() draw, if would find that no _NEW_POLYGONSTIPPLE state set,
thus will not call upload_polygon_stipple function.
This would fix all the oglc tristip test fail:
basic.strip
basic.fan
basic.triangles
Signed-off-by: Yuanhan Liu <yuanhan....@linux.intel.com>
---
src/mesa/drivers/dri/i965/brw_state_upload.c | 11 +++++++++++
1 files changed, 11 insertions(+), 0 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_state_upload.c
b/src/mesa/drivers/dri/i965/brw_state_upload.c
index 76ffa0d..e34f190 100644
--- a/src/mesa/drivers/dri/i965/brw_state_upload.c
+++ b/src/mesa/drivers/dri/i965/brw_state_upload.c
@@ -467,6 +467,17 @@ void brw_validate_state( struct brw_context *brw )
brw->state.dirty.brw |= BRW_NEW_VERTEX_PROGRAM;
}
+ /*
+ * Make sure _NEW_POLYGONSTIPPLE is set
+ *
+ * A glClear() call after glPolygonStipple() would reset
+ * this state. This would cause the later _real_ draw
+ * submitted with no polygon stipple info(will not call
+ * upload_polygon_stipple).
+ */
+ if (ctx->Polygon.StippleFlag)
+ ctx->NewState |= _NEW_POLYGONSTIPPLE;
+
if ((state->mesa | state->cache | state->brw) == 0)
return;
--
1.7.4.4
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
